If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to Atsushi Suzuki, Ph.D., Division of Organogenesis and Regeneration, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.
Snail is a transcription factor that regulates many cellular events involved in development, homeostasis, and disease. In hepatocellular carcinoma (HCC), Snail induces epithelial-to-mesenchymal transition that confers invasive properties on tumor cells during HCC progression and malignancy. Snail activation observed in HCC mouse models suggests its involvement not only in progression, but also onset of HCC. However, it remains unclear whether Snail directly contributes to HCC initiation or whether it supports HCC initiation promoted by other oncogenes. In this study, we generated mouse models for liver-specific and hepatocyte-specific overexpression of Snail to show the independent roles of Snail in liver homeostasis and disease. Enforced Snail expression resulted in liver and hepatocyte enlargement, inflammatory cell infiltration in the liver, lipid accumulation in hepatocytes, substantial increases in serum alanine aminotransferase and bile acids, yellow discoloration of tissues caused by bilirubin accumulation, and liver tumorigenesis. Snail overexpression suppressed mRNA expression of the tight junction components claudins and occludin and that of proteins associated with bile acid metabolism, leading to disruption of the biliary canaliculus formed among hepatocytes and excretion of abnormal amounts of unusual bile acids from hepatocytes. In conclusion, enforced Snail expression in hepatocytes is sufficient for induction of steatohepatitis and liver tumorigenesis through disruption of the biliary canaliculus and bile acid homeostasis in the liver.
Snail (also known as Snai1 and Sna) is a zinc-finger transcription factor involved in many aspects of cellular processes, including cell motility, differentiation, proliferation, and survival, as well as in embryonic development, tissue homeostasis, and disease.
For example, Snail expression normally is repressed during kidney development because Snail disrupts epithelial homeostasis in the kidney by repressing Cadherin-16 expression,
During skin development, Snail is expressed specifically in the hair buds, and sustained expression of Snail in the adult mouse skin results in epidermal hyperproliferation and differentiation defects.
Intestinal epithelial cells that compose the adult mouse intestines also express Snail, and abnormal expression levels of Snail in the intestinal epithelium lead to disruption of intestinal homeostasis.
Furthermore, pathologic activation of Snail in various types of epithelial tumors can direct tumor progression by inducing epithelial-to-mesenchymal transition and loss of cell adhesion, leading to the acquisition of invasive properties.
and glycogen synthase kinase-3β–dependent Snail degradation acts as a fundamental cue for initiation of hepatocyte proliferation during liver regeneration after transient acute liver failure.
Increased Snail activity also is related to hepatocellular carcinoma (HCC) progression and malignancy through down-regulation of E-cadherin (E-cad) and induction of epithelial-to-mesenchymal transition and invasive properties in tumor cells.
In an HRASG12V-driven HCC mouse model, transforming growth factor-β–induced Snail activation under suppression of p53 or activation of transcriptional coactivator with PDZ domain–binding motif promoted liver tumorigenesis,
suggesting that Snail is involved not only in progression, but also initiation, of HCC. However, it remains unknown whether Snail acts as a central player for HCC induction or as a promoter of HCC initiation induced by other oncogenes. Moreover, the underlying mechanism for HCC development in response to Snail activation remains unclear.
To address these issues, Snail was activated in the liver without any oncogenic events and the independent roles of Snail in liver homeostasis and disease were studied. Our data show that enforced Snail expression in hepatocytes is sufficient for induction of steatohepatitis and liver tumorigenesis in the absence of any additional cues for HCC development by disrupting the biliary canaliculus and bile acid homeostasis in the liver. These results suggest that Snail activation and Snail-mediated disruption of the biliary canaliculus and bile acid homeostasis in the liver act as an inducer of liver injury and contribute to the onset and progression of liver disease.
Materials and Methods
Mice
C57BL/6 mice (Clea Japan, Tokyo, Japan), Alb-Cre mice
(a gift from Frank Costantini, Columbia University, New York, NY) were used. In addition, a CAG-floxed neo-Snail mouse strain was generated to induce Cre-mediated overexpression of a mouse Snail full-length cDNA tagged with a hemagglutinin (HA) epitope
(a gift from Jun-ichi Miyazaki, Osaka University, Osaka, Japan) was used to drive transgene expression. For induction of Cre activity, 10-week–old Alb-CreERT2;CAG-floxed neo-Snail mice were given a single subcutaneous injection of tamoxifen (TM) (7.5 mg/mouse; Sigma-Aldrich, St. Louis, MO) dissolved in olive oil (Nacalai Tesque, Kyoto, Japan) at a concentration of 50 mg/mL. The experiments were approved by the Kyushu University Animal Experiment Committee, and the care and use of the animals were performed in accordance with institutional guidelines.
Immunostaining
For frozen tissue sections, liver tissues were embedded directly in Tissue-Tek OCT optimal cutting temperature compound (Sakura Finetek Japan, Tokyo, Japan) and sectioned. The frozen tissue sections were fixed with 4% paraformaldehyde for 5 minutes at room temperature. For paraffin-embedded tissue sections, liver tissues were fixed with Zinc Formalin Fixative (Polysciences, Warrington, PA), dehydrated in ethanol and xylene, embedded in paraffin wax, and sectioned. After deparaffinization and rehydration of the sections, antigen retrieval was performed by microwaving in 0.01 mol/L citrate buffer (pH 6.0). For immunohistochemistry, the sections were incubated with 0.3% hydrogen peroxide in methanol for 20 minutes at room temperature to quench endogenous peroxidase activity. After washing in phosphate-buffered saline and blocking with Block Ace (DS Pharma Biomedical, Osaka, Japan), the sections were incubated with the following primary antibodies: mouse anti-HA (1:100; Covance, Princeton, NJ); mouse anti–hepatocyte nuclear factor 4α (Hnf4α; 1:500; PPMX, Tokyo, Japan); mouse anti–E-cad (1:1000; BD Biosciences, San Jose, CA); mouse anti–N-cadherin (1:1000; BD Biosciences); rabbit anti-cytokeratin 19
(1:2000); rabbit anti–claudin-3 (Cldn3) (1:300; Abcam, Cambridge, MA); mouse anti–carcinoembryonic antigen–related cell adhesion molecule 1 (1:500; eBioscience, San Diego, CA); mouse anti–β-catenin (1:500; BD Biosciences); rat anti-CD45 (1:500; Novus Biologicals, Centennial, CO); rabbit anti–phosphorylated histone H3 (1:500; Millipore, Burlington, MA); and goat anti–green fluorescent protein/yellow fluorescent protein (1:2000; Abcam). After washing, the sections were incubated with a horseradish peroxidase–conjugated goat anti-rabbit IgG antibody (1:1000; Dako, Carpinteria, CA) for immunohistochemistry or with Alexa 488– and Alexa 555–conjugated secondary antibodies (1:1000; Thermo Fisher Scientific, Waltham, MA) specific to the species of the primary antibodies with DAPI for immunofluorescence staining.
Gene Expression Analysis
Total RNA was prepared from the liver tissues using a Nucleospin RNA II Kit (Takara Bio, Shiga, Japan), according to the manufacturer’s instructions, and cDNAs were synthesized from the total RNA as described.
TaqMan probes for Snail (Mm00441533_g1), occludin (Ocln) (Mm00500912_m1), and Cldn3 (Mm00515499_s1) were used. The TaqMan gene expression assay IDs (Applied Biosystems) are shown in parentheses after the gene names. As a normalization control, TaqMan rodent glyceraldehyde-3-phosphate dehydrogenase control reagents (Applied Biosystems) were used. To examine the expression of Cyp8b1, Cyp7b1, and Hsd3b7, qPCR was conducted using THUNDERBIRD SYBR qPCR Mix (Toyobo, Osaka, Japan), according to the manufacturer's instructions. qPCR primers for Cyp8b1 (forward: 5′-TGGAAGCCAAGAAGTCGTTC-3′ and reverse: 5′-TCCTCCTGTACCACCCTGAG-3′), Cyp7b1 (forward: 5′-CAATGACCCGGAAATCTTCGA-3′ and reverse: 5′-AGCTTCTCTGCCACACTTTCA-3′), Hsd3b7 (forward: 5′-ATGGTGAAGGTCATCAGGTCATGA-3′ and reverse: 5′-CCAGTATGTGCATCCAAGCAACAT-3′), and GAPDH (forward: 5′-TGTGTCCGTCGTGGATCTGA-3′ and reverse: 5′-TTGCTGTTGAAGTCGCAGGAG-3′) were used. The values for GAPDH were used as normalization controls.
Western Blot Analysis
Western blot analyses were conducted as described previously.
The following primary antibodies were used: goat anti-Snail (1:200; Santa Cruz Biotechnology, Santa Cruz, CA); mouse anti-HA (1:200; Covance); and mouse anti–β-actin (1:5000; Abcam).
Measurements of Total Bile Acid, Bilirubin, Alkaline Phosphatase, Alanine Aminotransferase, and Aspartate Aminotransferase
Mouse serum was collected from the femoral artery. The amounts of total bile acid, bilirubin, and alkaline phosphatase in the serum were measured using a Total Bile Acids Assay Kit (Diazyme Laboratories, Poway, CA), QuantiChrom Bilirubin Assay Kit (Bioassay Systems, Hayward, CA), and LabAssay alkaline phosphatase Kit (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), respectively. The amounts of alanine aminotransferase and aspartate aminotransferase in the serum were measured using a Transaminase CII Test Kit (FUJIFILM Wako Pure Chemical Corporation).
Paraffin-embedded liver tissue sections were stained using an Apoptosis In Situ Detection Kit (FUJIFILM Wako Pure Chemical Corporation), according to the manufacturer's instructions.
Gene Expression Microarray and Data Analysis
Total RNA was extracted from the liver of 13-week–old wild-type and Alb-Cre;CAG-floxed neo-Snail mice using an RNeasy Mini Kit (Qiagen, Hilden, Germany). Microarray analyses were performed as described previously (Cell Innovator, Fukuoka, Japan).
Functional enrichment analyses of the differentially expressed genes were performed using the Database for Annotation, Visualization, and Integrated Discovery.
Frozen sections of liver tissues were fixed with 4% paraformaldehyde for 5 minutes at room temperature, washed with phosphate-buffered saline, and incubated with 60% Oil red O (Muto Pure Chemicals, Tokyo, Japan) in water for 30 minutes at room temperature. After washing with 60% isopropanol for 1 minute and additional washing with phosphate-buffered saline, the sections were incubated with hematoxylin (Muto Pure Chemicals) for 5 minutes.
Electron Microscopy and Analysis of the Serum Bile Acid Composition
Transmission electron microscopy and scanning electron microscopy were performed by BML (Tokyo, Japan) and the Laboratory for Research Support, Medical Institute of Bioregulation (Kyushu University, Fukuoka, Japan), respectively. Bile acid composition in mouse serum was analyzed by the Junshin Clinic Bile Acid Institute (Tokyo, Japan).
Statistical Analysis
Statistical significance was analyzed using an unpaired t-test. Kaplan–Meier survival curves were analyzed statistically by the log-rank test. Differences at P < 0.05 were considered statistically significant.
Results
Generation of Transgenic Mice Overexpressing Snail in the Liver
To achieve overexpression of Snail in the liver, a mouse strain expressing Cre recombinase under the control of the mouse albumin enhancer/promoter (Alb-Cre mice)
was crossed with a CAG-floxed neo-Snail mouse strain. The CAG-floxed neo-Snail mice were generated in this study to induce Cre-mediated overexpression of an HA-tagged mouse Snail full-length cDNA (Figure 1A). As expected, the double-transgenic mice, such as Alb-Cre;CAG-floxed neo-Snail mice, expressed a large amount of Snail mRNA and Snail protein in the liver (Figure 1, B and C), and the HA tag was observed in Hnf4α-positive hepatocytes (Figure 1D). The liver zonation in the expression patterns of E-cad and N-cadherin indicated specific expression of E-cad and N-cadherin in hepatocytes residing around the portal and central veins, respectively (Figure 1E). In the liver of Alb-Cre;CAG-floxed neo-Snail mice, the number of E-cad–positive hepatocytes was decreased, and N-cadherin–positive hepatocytes replaced E-cad–positive hepatocytes in the periportal zone of the hepatic lobule (Figure 1E). These data show that Snail down-regulated the expression of its direct target gene E-cad in periportal hepatocytes. In addition to the effect of Snail overexpression on hepatocytes, the number of cholangiocytes composing the intrahepatic bile ducts was decreased significantly in Alb-Cre;CAG-floxed neo-Snail mice (Figure 1F). Global gene expression profiles in the liver of wild-type and Alb-Cre;CAG-floxed neo-Snail mice showed different gene expression patterns in both types of livers, including reduction of the cholangiocyte markers cytokeratin 19, Klf5, Prom1 (also known as CD133), and Onecut1 (also known as Hnf6) in the liver of Alb-Cre;CAG-floxed neo-Snail mice (Supplemental Figure S1, A and B). Because Cre recombinase is expressed not only in mature hepatocytes, but also bipotent hepatoblasts during liver development, cholangiocytes differentiated from hepatoblasts also expressed the transgenes, as well as hepatocytes (Supplemental Figure S2). Thus, the decrease in the number of cholangiocytes in the liver of Alb-Cre;CAG-floxed neo-Snail mice suggests that Snail activation inhibits cholangiocyte differentiation from hepatoblasts.
Figure 1Overexpression of the Snail gene in the mouse liver. A: Experimental procedure for overexpression of a Snail-HA fusion gene in specific cell types. After cell-specific Cre-mediated recombination, the loxP-flanked neo-stop cassette is removed, leading to the expression of the Snail-HA fusion gene under the regulation of the CAG promoter. B: Quantitative PCR analyses of Snail were performed on total RNA obtained from the livers of 8-week–old wild-type (WT) (1) and Alb-Cre;CAG-floxed neo-Snail (2) mice. All data were normalized by the values in the livers of WT mice, and the fold differences are shown. C: Western blot analyses were conducted for the livers of 8-week–old WT (1) and Alb-Cre;CAG-floxed neo-Snail (2) mice, and representative data are shown. β-actin was used for normalization. D: Co-immunofluorescence staining for hemagglutinin (HA) and hepatocyte nuclear factor 4α (Hnf4α) in the livers of 4-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative fluorescence images of Hnf4α-positive hepatocytes are shown. DNA was stained with DAPI. E: Co-immunofluorescence staining for N-cadherin (N-cad) and E-cadherin (E-cad) in the livers of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative fluorescence images of the periportal (top row) and pericentral (bottom row) zones of the hepatic lobule are shown. DNA was stained with DAPI. F: Immunohistochemical staining of the cholangiocyte marker cytokeratin 19 (CK19) in the livers of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative images are shown. The graph shows the percentages of CK19-positive cells residing around each portal vein (PV) in the livers of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Ten PVs were selected randomly in three tissue sections from the liver of each mouse. Data are expressed as means ± SD. n = 3 (B and F). ∗∗P < 0.01. Scale bars: 20 μm (D); 100 μm (E and F). CV, central vein.
Liver-Specific Snail Overexpression Induces Steatohepatitis and Lethal Tumorigenesis
The liver of Alb-Cre;CAG-floxed neo-Snail mice became massively enlarged compared with that of wild-type mice, regardless of age or sex, and small tumors developed in the liver of the double-transgenic mice at approximately 24 weeks after birth, and subsequently grew to form large tumors (Figure 2, A and B ). Survival curves showed that both male and female Alb-Cre;CAG-floxed neo-Snail mice died earlier than wild-type mice and that no double-transgenic mice survived until 80 weeks after birth (Figure 2C). Histologic analyses showed that the hepatocytes in Alb-Cre;CAG-floxed neo-Snail mice increased in size and accumulated large amounts of lipids in their cytoplasm, and that inflammatory cell infiltration occurred in the Snail-induced fatty liver (Figure 2D and Supplemental Figure S3A). Consistent with these data, gene ontology enrichment analysis showed that the extracted genes whose expression levels were significantly different between the liver of wild-type and Alb-Cre;CAG-floxed neo-Snail mice were highly enriched in genes associated with lipid metabolism (Supplemental Figure S1A). Although hepatic lipid accumulation was observed uniformly in the liver, hepatocyte enlargement initially occurred in the pericentral zone and gradually expanded to the entire hepatic region (Figure 2D and Supplemental Figure S3B). Meanwhile, cell proliferation and apoptosis were not induced in the liver of Alb-Cre;CAG-floxed neo-Snail mice (Supplemental Figure S3, C and D), suggesting that the liver hypertrophy in these mice arose from an increase in hepatocyte size. Taken together, these data show that enforced expression of Snail in the liver can induce steatohepatitis and liver tumorigenesis independently in the absence of any oncogenic stimuli.
Figure 2Snail overexpression in the mouse liver causes hepatomegaly, steatohepatitis, and liver tumorigenesis. A: Representative liver morphologies in 4-week–old (4w), 8w, 12w, 24w, 36w, and 48w wild-type (WT) and Alb-Cre;CAG-floxed neo-Snail mice. Arrowheads indicate tumors formed in the livers of Alb-Cre;CAG-floxed neo-Snail mice. B: Liver-to-body-weight ratios in male and female 4w, 8w, 12w, 24w, 36w, and 48w WT and Alb-Cre;CAG-floxed neo-Snail mice. C: Kaplan–Meier survival curves of male and female WT and Alb-Cre;CAG-floxed neo-Snail mice. There is no significant difference between male and female WT mice (P = 0.5432) or between male and female Alb-Cre;CAG-floxed neo-Snail mice (P = 0.0528). D: Representative images of hematoxylin and eosin (H&E)-stained and Oil red O–stained liver tissues obtained from 8w WT mice and 4w, 8w, 12w, 24w, and 36w Alb-Cre;CAG-floxed neo-Snail mice. Arrowheads indicate a representative area of inflammatory cell infiltration. Data are expressed as means ± SD (B). n = 3 (B). ∗P < 0.05, ∗∗P < 0.01. Scale bars: 1 cm (A); 100 μm (D, H&E); 200 μm (D, Oil red O). CV, central vein; PV, portal vein.
Snail Overexpression in the Liver Results in a Defect of Biliary Excretion
In the serum of Alb-Cre;CAG-floxed neo-Snail mice, the amounts of alanine aminotransferase, a sensitive indicator of liver damage, and of total bile acid and bile pigment bilirubin were increased significantly compared with those in wild-type mice (Figure 3, A–C). The accumulation of bilirubin in the blood and tissues of Alb-Cre;CAG-floxed neo-Snail mice resulted in yellow discoloration of their tissues, known as jaundice or hyperbilirubinemia (Figure 3D). Moreover, Alb-Cre;CAG-floxed neo-Snail mice excreted abnormal amounts of unusual bile acids in the urine (Table 1), and some genes involved in bile acid synthesis were down-regulated in their livers (Figure 3E). In contrast, the expression levels of genes encoding bile acid transporters and nuclear receptors were increased or invariable, respectively, in the livers of Alb-Cre;CAG-floxed neo-Snail mice compared with that of wild-type mice (Supplemental Figure S1B). Interestingly, as shown in Figure 2D, the hepatocyte enlargement and hepatic lipid accumulation observed in Alb-Cre;CAG-floxed neo-Snail mice were more severe in the livers of 8-week–old mice compared with 4-week–old mice. Indeed, the amounts of serum alanine aminotransferase in Alb-Cre;CAG-floxed neo-Snail mice at 8 weeks after birth were substantially higher than those at 4 weeks after birth (Figure 3A). However, significant increases in serum total bile acid and bilirubin already were detected in 4-week–old Alb-Cre;CAG-floxed neo-Snail mice (Figure 3, B and C), in which the degree of liver damage remained relatively low. Because an increase in bile acids is a cue for liver injury,
our data suggest that defective biliary excretion occurred before the development of hepatic damage in the liver of young Alb-Cre;CAG-floxed neo-Snail mice and became a key effector for induction of subsequent liver injury.
Figure 3Liver-specific Snail overexpression induces abnormal biliary excretion and liver injury. A–C: Amounts of alanine aminotransferase (ALT) (A), total bile acid (TBA) (B), and bilirubin (C) in the serum of 4-week–old (4w) and 8w wild-type (WT) and Alb-Cre;CAG-floxed neo-Snail mice. D: Representative images of subcutaneous and visceral tissues and serum of 12-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. E: Quantitative PCR analyses of total RNA obtained from the liver of 8w WT and Alb-Cre;CAG-floxed neo-Snail mice. All data were normalized by the values for the livers of WT mice, and the fold differences are shown. Red lines indicate means. Data are expressed as means ± SD (A–C and E). n = 3 (A–C); n = 10 WT mice (E); n = 7 Alb-Cre;CAG-floxed neo-Snail mice (E). Scale bars = 1 cm. ∗P < 0.05, ∗∗P < 0.01.
Hepatocyte-Specific Snail Overexpression Is Sufficient for Induction of Steatohepatitis and Liver Tumorigenesis
The number of cholangiocytes was decreased markedly in the liver of Alb-Cre;CAG-floxed neo-Snail mice (Figure 1F). Thus, the defect in biliary excretion in these mice may arise from a significant decrease in intrahepatic bile ducts and induce Snail-dependent liver injury. To examine this possibility, a mouse strain expressing an inducible form of Cre recombinase (CreERT2) from the albumin genomic locus (Alb-CreERT2 mice)
was crossed with the CAG-floxed neo-Snail mouse strain generated in this study. In the double-mutant mice, administration of TM allowed hepatocyte-specific overexpression of Snail. The intrahepatic bile ducts in the TM-administered Alb-CreERT2;CAG-floxed neo-Snail mice appeared normal compared with those in wild-type mice (Figure 4, A and B ). However, other phenotypic characteristics found in TM-administered Alb-CreERT2;CAG-floxed neo-Snail mice were similar to those in Alb-Cre;CAG-floxed neo-Snail mice, such as hepatocyte and liver enlargement accompanied by inflammatory reaction, hepatic lipid accumulation, and liver tumorigenesis (Figure 4, A and C). Moreover, the amounts of alanine aminotransferase, alkaline phosphatase, aspartate aminotransferase, total bile acid, and bilirubin in the serum of TM-administered Alb-CreERT2;CAG-floxed neo-Snail mice were increased substantially (Figure 4D). These data show that defective biliary excretion and liver injury result from a property change in hepatocytes, rather than in cholangiocytes, as a result of Snail overexpression.
Figure 4Phenotypic alterations in mice overexpressing Snail in the liver arise from a property change in hepatocytes, rather than in cholangiocytes. A: The livers were obtained from tamoxifen (TM)-administered 26-week–old wild-type (WT) mice and TM-administered 26-week–old and 72-week–old Alb-CreERT2;CAG-floxed neo-Snail mice. Representative liver morphologies, representative images of hematoxylin and eosin (H&E)-stained and Oil red O–stained liver tissues, and representative immunohistochemical images of cytokeratin 19 (CK19) are shown. B: Percentages of cells immunoreactive for CK19 residing around each portal vein (PV) in the livers of TM-administered 26-week–old WT (1) and Alb-CreERT2;CAG-floxed neo-Snail (2) mice. Ten PVs were selected randomly in three tissue sections from the liver of each mouse. C: Liver-to-body-weight ratios in TM-administered 26-week–old WT (1) and Alb-CreERT2;CAG-floxed neo-Snail (2) mice. D: Amounts of alanine aminotransferase (ALT), alkaline phosphatase (ALP), aspartate aminotransferase (AST), total bile acid (TBA), and bilirubin in the serum of TM-administered 26-week–old WT (1) and Alb-CreERT2;CAG-floxed neo-Snail (2) mice. Data are expressed as means ± SD. n = 3 (B–D). Scale bars: 1 cm (liver images); 100 μm (liver tissue sections). ∗P < 0.05, ∗∗P < 0.01.
Disruption of the Biliary Canaliculus as a Result of Snail Overexpression in Hepatocytes
Next, the changes that occurred in the Snail-overexpressing hepatocytes were studied for induction of the defect in biliary excretion during the early stage of liver injury. Previous studies showed that Snail directly suppresses mRNA expression of the tight junction (TJ) components claudins and Ocln.
Consistently, our qPCR analyses showed that the expression levels of Cldn3 and Ocln in the liver of Alb-Cre;CAG-floxed neo-Snail mice were much lower than the levels in wild-type mice (Figure 5A). Furthermore, immunofluorescence analyses showed the absence of Cldn3 and abnormal localization of anti–carcinoembryonic antigen–related cell adhesion molecule 1 (also known as biliary glycoprotein) normally localized at hepatocyte TJs with Cldn3, in the liver of not only adult (8-week–old), but also young (4-week–old), Alb-Cre;CAG-floxed neo-Snail mice (Figure 5B). In addition to Cldn3 and Ocln, the expression levels of genes encoding other TJ components also were down-regulated in the liver of Alb-Cre;CAG-floxed neo-Snail mice (Supplemental Figure S1C). Because hepatocytes form the biliary canaliculus by adhering to adjacent hepatocytes through TJs, our data suggest that Snail-induced suppression of the integral membrane proteins localized at TJs disrupted the structure of the biliary canaliculus and led to the defect in biliary excretion. In fact, transmission electron microscopy and scanning electron microscopy observations showed that the biliary canaliculus disappeared from the liver of Alb-Cre;CAG-floxed neo-Snail mice through disruption of hepatocyte TJs (Figure 5C). The resulting extension of the intercellular space formed between hepatocytes in the liver of Alb-Cre;CAG-floxed neo-Snail mice may allow bile acid diffusion into the blood circulatory system. In addition to the structural defect in hepatocyte TJs, ultrastructural analyses showed characteristic mitochondrial abnormalities in hepatocytes of Alb-Cre;CAG-floxed neo-Snail mice, suggesting induction of oxidative stress based on an increase in reactive oxygen species production for liver damage (Figure 5C). Disruption of the hepatocyte TJs, characterized by loss of Cldn3 and abnormal localization of anti–carcinoembryonic antigen–related cell adhesion molecule 1 in the liver, also was observed in TM-administered Alb-CreERT2;CAG-floxed neo-Snail mice (Figure 5D). Taken together, our data show that enforced Snail expression in hepatocytes causes disruption of the biliary canaliculus through suppressed expression of TJ components and leads to a defect in biliary excretion from an early stage of liver injury.
Figure 5Snail overexpression in the mouse liver causes disruption of the biliary canaliculus formed among hepatocytes. A: Quantitative PCR analyses on total RNA obtained from the livers of 8-week–old (8w) wild-type (WT) and Alb-Cre;CAG-floxed neo-Snail mice. All data were normalized by the values for the livers of WT mice, and the fold differences are shown. B: Co-immunofluorescence staining for claudin-3 (Cldn3) and anti–carcinoembryonic antigen–related cell adhesion molecule 1 (CEACAM-1) in the livers of 4w and 8w WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative fluorescence images are shown. DNA was stained with DAPI. C: Representative transmission electron microscopy (TEM) and scanning electron microscopy (SEM) images of liver tissues obtained from 8w WT and Alb-Cre;CAG-floxed neo-Snail mice. Arrowheads indicate the tight junction and biliary canaliculus formed among hepatocytes, respectively, both of which were observed in the livers of WT mice, but not in the livers of Alb-Cre;CAG-floxed neo-Snail mice. Instead, many lipid droplets and abnormal mitochondria were observed within hepatocytes in the livers of Alb-Cre;CAG-floxed neo-Snail mice. D: Co-immunofluorescence staining for Cldn3 and CEACAM-1 in the livers of TM-administered 22-week–old WT and Alb-CreERT2;CAG-floxed neo-Snail mice. Representative fluorescence images are shown. DNA was stained with DAPI. Data are expressed as means ± SD. n = 3 (A). Scale bars: 20 μm (B and D); 10 μm (C). ∗P < 0.05.
In this study, we showed that enforced Snail expression in hepatocytes is able to induce steatohepatitis and liver tumorigenesis independently through disruption of the biliary canaliculus and bile acid homeostasis in the liver. Thus, Snail may act not only as an epithelial-to-mesenchymal transition inducer in HCC for tumor progression and invasion,
but also a key player for HCC induction. Our findings suggest that Snail activation and Snail-mediated disruption of the biliary canaliculus and bile acid homeostasis in the liver are involved in the pathogenesis of liver disease. Indeed, Snail up-regulation is detected in mouse hepatocytes responding to chronic hepatic injury, leading to progression of liver fibrosis,
In this study, the pathologic changes found in the liver of both Alb-Cre;CAG-floxed neo-Snail mice and Alb-CreERT2;CAG-floxed neo-Snail mice, such as fatty liver and subsequent steatohepatitis, appear similar to the signs and symptoms of chronic liver disease such as alcoholic and nonalcoholic fatty liver disease.
Thus, it can be suggested that transforming growth factor-β–induced Snail activation and Snail-induced Ocln down-regulation occur in response to alcohol intake and increase susceptibility to alcohol-induced hepatic injury.
TJs that produce hepatocyte polarity in the liver and form the biliary canaliculus structure among hepatocytes are composed of the integral membrane proteins claudins, Ocln, and junctional adhesion molecules, and many peripheral membrane proteins, including zonula occludens-1, -2, and -3, which bind intracellularly to claudins, Ocln, and junctional adhesion molecules.
Enforced Snail expression in hepatocytes is sufficient for disruption of the TJs formed among hepatocytes. Because Snail directly suppresses the expression of claudins and Ocln,
it is suggested that the disruption of hepatocyte TJs in the liver of both Alb-Cre;CAG-floxed neo-Snail mice and Alb-CreERT2;CAG-floxed neo-Snail mice results from significant decreases in the TJ components claudins and Ocln. However, because TJs are composed of many proteins, it is difficult to conclude that only decreases in claudins and Ocln within hepatocytes disrupt the formation and maintenance of TJs. Indeed, TJ components other than claudins and Ocln also were reduced in the liver of Alb-Cre;CAG-floxed neo-Snail mice. Thus, it is suggested that Snail directly or indirectly can suppress the mRNA expression of TJ components, and thus the expression levels of Snail in hepatocytes may be important for maintenance of the physiological function of hepatocyte TJs in the liver. In addition to disruption of the biliary canaliculus, enforced Snail expression in the liver results in down-regulated expression of genes associated with bile acid synthesis, such as Cyp8b1, Cyp7b1, and Hsd3b7, and induces excretion of abnormal amounts of unusual bile acids from hepatocytes. Human CYP8B1, CYP7B1, and HSD3B7 are the genes responsible for 12α-hydroxylase deficiency, oxysterol 7α-hydroxylase deficiency, and 3β-hydroxy-Δ5-C27-steroid dehydrogenase/isomerase deficiency, respectively, which are observed in diseases involving inborn errors of bile acid metabolism.
Thus, Snail contributes to the maintenance of bile acid homeostasis by regulating the expression levels of key genes involved in bile acid metabolism in hepatocytes.
Bile acids induce mitochondrial dysfunction, oxidative stress, and reactive oxygen species generation in hepatocytes,
Thus, increased and abnormal biliary excretion as a result of Snail-induced disruption of the biliary canaliculus and bile acid homeostasis in the liver may lead to hepatic damage mediated by oxidative stress through bile acid–induced mitochondrial alterations in hepatocytes. Indeed, the hepatocytes in Alb-Cre;CAG-floxed neo-Snail mouse livers contain many morphologically abnormal mitochondria. In the human fatty liver, the number of hepatocytes containing abnormal mitochondria is significantly higher than that in the normal liver.
Thus, it has been suggested that the retention of large amounts of bile acids in Alb-Cre;CAG-floxed neo-Snail mice and Alb-CreERT2;CAG-floxed neo-Snail mice allows the induction of liver tumorigenesis. Although bile acids induce apoptosis of hepatocytes,
the present data showed that an increase in bile acids caused by enforced Snail expression in the liver did not affect the induction of hepatocyte apoptosis. These results suggest that Snail suppresses the expression of genes not only for TJ components and bile acid synthesis, but also apoptosis in hepatocytes. Indeed, Snail is known to have a role in resistance to apoptosis,
as well as many other functions. The present study shows that the amount of Snail in hepatocytes is deeply related to the function of hepatocytes and should be controlled properly for the maintenance of liver homeostasis. Suppression of Snail up-regulation in hepatocytes that reside in the damaged liver may contribute to the prevention of liver disease progression.
Acknowledgments
We thank Drs. Pierre Chambon, Daniel Metzger, Frank Costantini, Antonio García de Herreros, and Jun-ichi Miyazaki for sharing reagents and providing mice; and Yuuki Honda, Chiaki Kaieda, Kanako Ichikawa, Ryo Ugawa, and Emiko Koba for excellent technical assistance.
Author Contributions
S.M. performed all experiments and collected and analyzed data; and A.S. conceived, designed, and managed the project and wrote the paper.
Supplemental Data
Supplemental Figure S1Global gene expression analyses of the liver of wild-type (WT) and Alb-Cre;CAG-floxed neo-Snail mice using a microarray. A: Heatmap showing genes with different expression levels between the livers of 13-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice (left panel). Gene ontology enrichment analysis was performed for genes that showed different expression levels between both types of livers (right panel). B: Heatmap showing the expression levels of genes encoding cholangiocyte markers [Klf5, Krt19 (cytokeratin 19 [CK19]), Onecut1, and Prom1], bile acid transporters (Abcb11, Abcc1, Abcc2, Abcc3, Abcc4, Abcc5, Abcc6, and Slc10a1), and nuclear receptors (Fxr1, Fxr2, Nr0b2, Nr5a2, Ppara, Rxra, Rxrg, and Vdr) in the livers of 13-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. C: Heatmap showing the expression levels of genes encoding tight junction components in the liver of 13-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice.
Supplemental Figure S2Hepatoblasts in Alb-Cre mice express Cre recombinase during liver development and give rise to both hepatocytes and cholangiocytes. The Alb-Cre mouse strain was crossed with the R26RYFP/YFP reporter mouse line. Co-immunofluorescence staining of yellow fluorescent protein (YFP) with cytokeratin 19 (CK19) was conducted for the livers of 10-week–old Alb-Cre;R26RYFP/+ mice. A representative fluorescence image of YFP-positive cholangiocytes is shown. Scale bars = 100 μm. PV, portal vein.
Supplemental Figure S3Additional analyses for the liver of Alb-Cre;CAG-floxed neo-Snail mice. A: Hematoxylin and eosin (HE) staining and immunofluorescence staining for CD45 were conducted in the livers of 36-week–old wild-type (WT) and Alb-Cre;CAG-floxed neo-Snail mice. Representative histologic and fluorescence images are shown. Arrowheads indicate representative areas of inflammatory cell infiltration. B: Immunofluorescence staining for β-catenin in the liver of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative fluorescence images are shown. The graphs show the percentages of β-catenin–positive cells residing around PVs and CVs in the livers of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Ten PVs and 10 CVs were selected randomly in three tissue sections from the liver of each mouse, and the number of β-catenin–positive cells were counted. C: Immunofluorescence staining for the mitosis-associated marker phosphorylated histone H3 (pHH3) in the liver of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice. Representative fluorescence images are shown. Arrowhead indicates a representative cell immunoreactive for pHH3. D: Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining in the liver of 8-week–old WT and Alb-Cre;CAG-floxed neo-Snail mice for detection of apoptotic cells. DNase I–treated liver tissues were used as positive controls (insets). Data are expressed as means ± SD. n = 3 (B). ∗P < 0.05. Scale bars: 100 μm (main images and insets). CV, central vein; PV, portal vein.
Supported in part by Japan Society for the Promotion of Science ( JSPS) KAKENHI grants 25713014 , JP16H01850 , JP17H05623 , JP17K19603 , JP18H05102 , JP19H01177 , and JP19H05267 ; the Core Research for Evolutional Science and Technology Program of the Japan Agency for Medical Research and Development ; the Program for Basic and Clinical Research on Hepatitis of the Japan Agency for Medical Research and Development ; the Practical Research Project for Rare/Intractable Diseases of the Japan Agency for Medical Research and Development ; the Research Center Network for Realization of Regenerative Medicine of the Japan Agency for Medical Research and Development ; the Takeda Science Foundation ; the Princess Takamatsu Cancer Research Fund ; and the Uehara Memorial Foundation .
30121-8&id=gr1.jpg){kind=link}
30121-8&id=gr2.jpg){kind=link}
30121-8&id=gr3.jpg){kind=link}
30121-8&id=gr4.jpg){kind=link}
30121-8&id=gr5.jpg){kind=link}
30121-8&id=figs1.jpg){kind=link}
30121-8&id=figs2.jpg){kind=link}
30121-8&id=figs3.jpg){kind=link}